Hey! I'm Gaurav. Thanks for visiting my corner of the internet! My interests lie at the nexus of product, behavioural science, and AI. <img src="h...
My favourite AI tools have one thing in common: they are low friction, high function . What does that mean? Low friction means designing interaction patterns between humans and LLMs in ways that require the fewest cognitive resources. High function means harnessing the growing inference capabilities of LLMs in ways that deliver outsized gains in productivity. Granola , an AI-powered notepad for…
Building an AI product with strong adoption and retention requires work on the wrapper, i.e. the app layer. The image on the left represents an AI product with a simple interface that interacts with industry-grade LLMs via straightforward API calls. The image on the right represents an AI product that dilutes its dependence on the API calls to LLMs, i.e. a product that's built by expending time…
Anthropomorphising LLMs is an active product choice. While reasoning models unlock superhuman capabilities, I believe the real winners in the agentic AI space will be the ones that succeed in making the models feel as human as possible. Each AI agent should take on a life of its own. The best human agents are the ones that cater their services to you . They adapt to evolving needs, they're always…
Humans build tools to make life easier. The wheel made it easier to get from point A to point B. The dishwasher made it easier to turn utensils from dirty to clean. The true value of a tool is measured by the question: "how much easier is my life with the presence of this tool?" Consider a lemon squeezer. Without it, you’re wrestling a lemon like it’s a stress ball. Juice sprays across the kitchen…
AI is not deterministic. It is stochastic. When you click a button in Excel, you expect the same result every time. This helps build confidence and trust in the tool. When you feed a prompt to GPT-5, it gives you slightly different answers every time. Sometimes brilliant. Sometimes bizarre. The next groundbreaking AI product will be one that focuses on minimizing the bizarre . Note that better…
Context windows are rising. Cost of inference is falling. What does this mean for products that are built on top of LLMs? I’ve been using Lovable to build web apps for the last 6 months. Pretty much ever since it launched. Recently, I was browsing through my prompt history to contextualize how rapidly LLMs become smarter. The first web app I built on Lovable was Expensify : a simple tool to track…
The golden rule of product is to enable users to complete their goals . Let's break this down. Users are humans, i.e complex arrangement of matter. Humans set goals, thanks to system 2 thinking. Humans fail to complete goals, thanks to system 1 thinking. A successful product must therefore prevent humans from failing to complete their goals. Here's somewhat of a playbook for building a successful…